16-883 Special Topics: Provably Safe Robotics

Time: Tuesday and Thursday 3:30 pm — 4:50 pm
Location: GHC 4102
Instructor: Changliu Liu (cliu6@andrew.cmu.edu)
Office Hours: Tuesday 2:30 pm — 3:30 pm in NSH 4525
Canvas: https://canvas.cmu.edu/courses/45724
Piazza: https://piazza.com/cmu/spring2025/16883

Course Description

Safe autonomy has become increasingly critical in many application domains. It is important to ensure not only the safety of the ego robot but also the safety of other agents (humans or robots) that directly interact with the autonomy. For example:

  • Robots should be safe for human workers in human-robot collaborative assembly.
  • Autonomous vehicles should be safe for other road participants.

For complex autonomous systems with many degrees of freedom, safe operation depends on the correct functioning of all system components (e.g., accurate perception, optimal decision making, and safe control). This course addresses both the design and verification of safe robotic systems, focusing on:

  • Designing safety through control and learning.
  • Verifying neural components, closed-loop systems, and safety/reachability analyses.

Learning Objectives

  • Familiarize with tools and methodologies for designing and verifying safe robotic systems.
  • Use these tools to improve or verify the safety of existing designs.
  • Develop new tools and methodologies for robot safety.

In this semeseter, we will focus on humanoid safety and be using SPARK for homework assignments.

Assessment Structure

  • Participation: 10 points
  • Homeworks: 45 points (3 assignments, 15 points each)
  • Project: 45 points
    • Proposal: 10 points
    • Presentation: 15 points
    • Final Report: 20 points

Policies

Class Attendance and Participation

Participation and attendance are critical. Notify the instructor beforehand if you must miss a class to arrange alternative ways to catch up.

Homeworks

Three homeworks will be assigned, each worth 15 points. You will have two weeks to complete each.

Project

  • Proposal (10 points): A 1-page abstract in IEEE format introducing the problem and outlining the plan for the project.

  • Final Report (20 points): A 6-page conference paper in IEEE format. The structure will differ based on paper type:
    • Survey Papers: Introduction, review of existing surveys, analysis/comparison, and discussion on learnings and future directions.
    • Technical Papers: Introduction/motivation, literature review, problem formulation/solution, results, and discussion.
  • Presentation (15 points): Each team will have 10–25 minutes for the final presentation, depending on team size. Presentations should address:
    • Major contributions.
    • Problem formulation.
    • Solution methodology.
    • Results.
    • Limitations and future work.
  • Projects can be individual or team-based (maximum of 3 members).

Extensions and Late Assignments

Late submissions are not accepted except in emergencies. Contact the instructor early to request extensions.

Weekly Schedule

Week 1

  • 1/14: Lecture 1 - Safety Specifications
  • 1/16: Lecture 2 - SPARK: Safe Humanoid Toolbox (Guest Lecture: Rui Chen)

Week 2

  • 1/21: Lecture 3 - Safe Control for Mobile Robots (Single Integrator)
  • 1/23: Lecture 4 - Safe Control for Mobile Robots (Double Integrator Unconstrained)

Week 3

  • 1/28: Lecture 5 - Safe Control for Mobile Robots (Double Integrator Constrained)
  • 1/30: Lecture 6 - Safe Control for Mobile Robots (Singularity)
  • 1/31: Homework 1 Released

Week 4

  • 2/4: Lecture 7 - Safe Control for Mobile Robots (Nonlinear Models)
  • 2/6: Lecture 8 - Safe Control for Mobile Robots (Hierarchical Models)

Week 5

  • 2/11: Lecture 9 - Safe Control for Manipulators (Planar Arms)
  • 2/13: Lecture 10 - Safe Control for Manipulators (3D Arms)
  • 2/14: Homework 1 Due; Homework 2 Released

Week 6

  • 2/18: Lecture 11 - Safe Control Theory (Forward Invariance)
  • 2/20: Lecture 12 - Safe Control Theory (Unified Framework)

Week 7

  • 2/25: Lecture 13 - Safe Control Theory (Synthesis)
  • 2/27: Lecture 14 - Safe Control Theory (Certificates)
  • 2/28: Homework 2 Due

Week 8: Spring Break

  • 3/4: No Class
  • 3/6: No Class

Week 9

  • 3/11: Lecture 15 - Safety Monitors for Data-Driven Control
  • 3/13: Lecture 16 - Reachability Analysis

Week 10

  • 3/18: Lecture 17 - Neural Certificates
  • 3/20: Lecture 18 - Certification of Neural Certificates
  • 3/21: Project Proposal Due; Homework 3 Released

Week 11

  • 3/25: Lecture 19 - Certification of Pose Estimation Models
  • 3/27: Lecture 20 - Certified Training

Week 12

  • 4/1: Lecture 21 - Safe Reinforcement Learning (GUARD)
  • 4/3: No Class
  • 4/4: Homework 3 Due

Week 13

  • 4/8: Lecture 22 - State-Wise Safe Reinforcement Learning
  • 4/10: Lecture 23 - Reactive Safety vs Long-Term Safety

Week 14

  • 4/15: Lecture 24 - Procedures for System Evaluation and Testing
  • 4/17: Lecture 25 - Safety and Responsibility in Human-Robot Interactions

Week 15

  • 4/22: Project Presentations
  • 4/24: Project Presentations

Week 16

  • 5/2: Project Report Due

Additional Policies and Resources

Academic Integrity

Plagiarism and cheating are strictly prohibited. Refer to CMU’s academic integrity policy for details. Cite all sources appropriately.

Accommodations for Students with Disabilities

Contact the Office of Disability Resources at access@andrew.cmu.edu for support and accommodations.

Student Support

If you experience stress or difficult life events, reach out to Counseling and Psychological Services (CaPS) or call 412-268-2922.